主权项 |
1. A method of managing backup meta data indexing in a backup process, the method comprising:
receiving, by a backup server, backup meta data from a client over a network, the backup meta data associated with a backup instance of the client, and wherein the backup meta data comprises one of directory data or node data; determining whether the backup meta data is directory data or node data, wherein directory data comprises a filename, an inode and a backup savetime identifying the backup instance, and node data comprises an inode and attributes of a node referenced by the inode; committing the backup meta data to a directory record of a directory record database associated with the client for storing directory data of the client in response to the backup meta data being directory data, wherein the backup server maintains a plurality of directory record databases corresponding to a plurality of clients, wherein each of the directory record databases includes a plurality of directory records corresponding to a plurality of directories or files of directories associated with the client; committing the backup meta data to a node record database for storing node data of the backup instance in response to the backup meta data being node data, wherein the node record database is one of a plurality of node record databases, each corresponding to one of a plurality of backup instances, wherein each node record database includes a plurality of node records, each node record representing one of a plurality of files associated with a corresponding backup instance, wherein the node records are indexed based on inodes of the files, wherein the directory record database is a separate database different from the node record database and at least one node record database is maintained per backup savetime; receiving a request from a first client for accessing a first backup instance of first backup data associated with the first client, wherein the request includes a first filename and a first savetime; in response to the request, identifying a first directory record database from the plurality of directory record databases based on a client identifier (ID) identifying the first client, the first directory record database including a plurality of directory records associated with the first client; obtaining a first node identifier associated with a first directory record comprising the first filename; identifying a first node record database based on the first savetime from the plurality of node record databases, the first node record database including a plurality of node records, each corresponding to a file or a directory of one or more files associated with the first backup instance; obtaining a first node record from the first node record database based on the first node identifier; and returning node record information associated with the first node record to the client. |